English

Low autocorrelated multi-phase sequences

Statistical Mechanics 2009-11-07 v1 Disordered Systems and Neural Networks

Abstract

The interplay between the ground state energy of the generalized Bernasconi model to multi-phase, and the minimal value of the maximal autocorrelation function, Cmax=maxKCKC_{max}=\max_K{|C_K|}, K=1,..N1K=1,..N-1, is examined analytically and the main results are: (a) The minimal value of minNCmax\min_N{C_{max}} is 0.435N0.435\sqrt{N} significantly smaller than the typical value for random sequences O(logNN)O(\sqrt{\log{N}}\sqrt{N}). (b) minNCmax\min_N{C_{max}} over all sequences of length N is obtained in an energy which is about 30% above the ground-state energy of the generalized Bernasconi model, independent of the number of phases m. (c) The maximal merit factor FmaxF_{max} grows linearly with m. (d) For a given N, minNCmaxN/m\min_N{C_{max}}\sim\sqrt{N/m} indicating that for m=N, minNCmax=1\min_N{C_{max}}=1, i.e. a Barker code exits. The analytical results are confirmed by simulations.

Cite

@article{arxiv.cond-mat/0103185,
  title  = {Low autocorrelated multi-phase sequences},
  author = {Liat Ein-Dor and Ido Kanter and Wolfgang KJinzel},
  journal= {arXiv preprint arXiv:cond-mat/0103185},
  year   = {2009}
}

Comments

4 pages, 4 figures